solution

Data integrity

In a Formula One Racetrack Project, carried out by IFM Sports Marketing Surveys, the data were obtained by face-to-face and telephone interviews. As the questionnaire was developed and finalised, a preliminary plan was drawn up of how the findings could be analysed. The questionnaires were edited by a supervisor as they were being returned from Australia, Brazil, France, Germany, Italy, Japan, Spain and the UK. The questionnaires were checked for incomplete, inconsistent and ambiguous responses. Questionnaires with problematic responses were queried with supervisors in each of the eight countries. In some circumstances, the supervisors were asked to re-contact the participants to clarify certain issues. Thirty-five questionnaires were discarded because the proportion of incomplete or unsatisfactory responses rendered them too poor to use. This resulted in a final sample size of 2,050.
A codebook was developed for coding the questionnaires; this was done automatically as the questionnaire was designed using the SNAP software (www.snapsurveys. com). Data were transcribed by being directly keyed in as the telephone interview mode was conducted. In the face-to-face mode, data were transcribed using personal digital assistants (PDAs) as the interviews were conducted. The software has a built-in error check that identified out-of-range responses; 10% of the data were verified for other data entry errors. The data were cleaned by identifying logically inconsistent responses. Most of the rating information was obtained using five-point scales, so responses of 0, 6 and 7 were considered out of range and a code of 9 was assigned to missing responses. If an out-of-range response was keyed in, the SNAP software issued an audible warning and prohibited any continuation of data entry. New variables were created that were composites of original variables. Finally, a data analysis strategy was developed.